A SIMULATION STUDY OF STATIC AND DYNAMIC DISTRIBUTED RESOURCE ALLOCATION ALGORITHMS

Abstract

Resource allocation is one of the prime issues in a distributed computer system. Resource deadlocks are traditional deadlocks in which all of some set of processes are blocked waiting for resources held by other blocked processes. In the present work two static message based protocols for solving this problem are studied in a distributed system. The first one requires the maintenance of a global queue and the second one uses many local virtual queues. The protocols make use of an exclusion graph, that is mapped on to a physical .set of processors and communication links, so that its edges may span zero or more communication links. The simulation studies showed that the first algorithm has better performance over the second algorithm, both in terms of mean message count and mean waiting time. A dynamic variation was applied to the first algorithm in order to make it work under a dynamic situation. In a dynamic system processes can change their resource requirements during execution. The variation also allows processes to be added and/or removed from the system.